From 515fb1318019e0121b8dcfa47ee996369fcb9f75 Mon Sep 17 00:00:00 2001
From: ClashSAN <98228077+ClashSAN@users.noreply.github.com>
Date: Sun, 27 Aug 2023 13:24:43 -0400
Subject: [PATCH] sd_xl_1.0_0.9vae model info is outdated/bad. Sorry.

---
 Features.md | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/Features.md b/Features.md
index 0f24508..bc1281b 100644
--- a/Features.md
+++ b/Features.md
@@ -14,11 +14,9 @@ All examples are non-cherrypicked unless specified otherwise.
 
 - [sd_xl_refiner_1.0_0.9vae](https://huggingface.co/stabilityai/stable-diffusion-xl-refiner-1.0/blob/main/sd_xl_refiner_1.0_0.9vae.safetensors)
 
-These models **specifically** are recommended for generating, merging and training.
+>~~They have a built-in trained vae by [madebyollin](https://github.com/madebyollin) which fixes NaN infinity calculations running in fp16. (Here is the most up-to-date [VAE](https://huggingface.co/madebyollin/sdxl-vae-fp16-fix/blob/main/sdxl_vae.safetensors) for reference)~~ Bad/Outdated info. Using this model will not fix fp16 issues for all models. You should merge this [VAE](https://huggingface.co/madebyollin/sdxl-vae-fp16-fix/blob/main/sdxl_vae.safetensors) with the models.
 
->They have a built-in trained vae by [madebyollin](https://github.com/madebyollin) which fixes NaN infinity calculations running in fp16. (Here is the most up-to-date [VAE](https://huggingface.co/madebyollin/sdxl-vae-fp16-fix/blob/main/sdxl_vae.safetensors) for reference)
-
-## SD-XL BASE
+# SD-XL BASE
 
 This is a model designed for generating quality `1024×1024`-sized images. It is **not** meant to generate good pictures at `512×512`.